Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already flag destructiveHint=true, idempotentHint=false and openWorldHint=true, but the description adds genuinely new facts: attendees receive a cancellation and lose the event from their calendars, the action is irreversible, and the return is the event with status cancelled. That is real side-effect disclosure beyond the structured fields.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Three short sentences, front-loaded with the core action, then the consequence, then the alternative. No filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a one-parameter mutation with full annotation coverage and no output schema, the description supplies the missing pieces an agent needs: irreversibility, attendee impact, and the returned status. Nothing material is absent.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100% and the single id parameter is already documented as coming from create_event or list_events. The description adds no format or sourcing detail beyond the schema, so the baseline of 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

Specific verb (cancel) plus resource (meeting) and the provenance constraint (created with create_event). It is unmistakably distinct from siblings like update_event and delete_message.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly routes the agent: use this to cancel, use update_event instead to move a meeting. The when-not condition is stated rather than inferred.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Adds substantial behavior beyond the annotations: the operation is a soft delete recoverable in Trash, it is gated on an operator-configured allow_delete flag, and the failure mode is the specific error code delete_not_allowed. This is valuable because destructiveHint=true alone would imply unrecoverable removal, and the description corrects that impression.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Three short sentences, no filler. The operation and its reversibility are front-loaded, followed by the precondition, the alternative, and the return shape in that order of importance.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a one-parameter destructive tool, everything an agent needs is present: what it does, the gating condition, the error code on refusal, the alternative tool, and the return value despite the absence of an output schema.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the single 'id' parameter is already documented as coming from list_messages. The description adds no further syntax, format, or validation detail beyond the schema, so the baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource+destination: 'Move a received message to the Trash folder'. The parenthetical clarifies the semantics of the destination (recoverable until the server empties it), which distinguishes this from a hard delete or from mark_message.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly names the alternative tool and the condition that selects it ('Use mark_message instead if you only want to mark it as handled') and states the enabling precondition (operator must have set allow_delete). Both the when and the when-not are covered.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already cover the safety profile (openWorld, non-idempotent, non-destructive). The description adds real behavioral value beyond them: the same allow list and review/send limit as send_message, the structure of the forwarded payload (note, original header/text, attachments), and the return shape.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Three sentences, front-loaded with the core action and followed by the alternative and constraints. Each sentence earns its place, though the parenthetical detail on attachments could be trimmed.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema, the description carries the return value ({ message_id, warnings }) and covers the key behavioral constraints. Complete enough for an agent to call it, though it omits details like attachment size or recipient limits.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the schema already documents all six parameters. The description restates the include_attachments default behavior and the body_markdown placement but adds little the schema doesn't already say, so the baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('Forward a received message to other allowed recipients') and names the sibling it is not (reply_message). An agent can distinguish this from send_message and reply_message without opening either schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Names the alternative explicitly ('Use reply_message to answer the sender instead') and gives context via the send_message constraints. It provides clear routing but stops short of spelling out full when/when-not conditions such as required permissions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint, idempotentHint and destructiveHint=false, so safety is covered structurally. The description still adds non-obvious behavior: attendee answers (accepted/declined/tentative) arrive automatically from participants' calendars rather than being set here. The closing 'changes nothing' is largely redundant with readOnlyHint.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Three short sentences, front-loaded with purpose and return contents, then the id-procurement hint. Nearly every clause earns its place, though 'changes nothing' duplicates the readOnly annotation.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema, the description compensates by enumerating the returned payload (times in UTC and local, place, attendees and answers), and it explains how to obtain the required id. Nothing an agent needs to invoke it correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Only one parameter and schema description coverage is 100%, so the schema already documents the id fully (including its origin). The description restates that origin ('Use list_events to find an id') without adding format, validation, or error semantics, which is the expected baseline when the schema does the work.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb (get) and resource (one meeting), then enumerates the returned fields: title, UTC/local times, place, attendees and their answer states. It explicitly ties the resource to create_event and names list_events as the id source, so an agent can distinguish it from the sibling list_events and update_event.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives clear operational context: the event must have been created with create_event, and list_events is the way to obtain an id. It does not spell out when to prefer this over a hypothetical alternative for a single event, but for a one-id read tool the guidance is effectively complete.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint, idempotentHint, destructiveHint=false and openWorldHint=false, so the safety profile is fully covered by structured data. The description's 'changes nothing' largely restates destructiveHint=false; the only genuine addition is that it returns a JSON object and should be called once per session.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences, no padding. The first front-loads what the tool returns; the second carries the actionable ordering guidance. Every clause earns its place.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema, the description compensates by enumerating the contents of the returned JSON, which is enough to set expectations. It stops short of describing the shape or fields of the policies/limits, but for a read-only, zero-parameter helper this is largely complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The tool takes zero parameters, so the schema has nothing to describe and there is no semantic gap for the description to fill. Baseline 4 applies for a parameterless tool.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description names a precise verb and enumerates the specific resources returned: the user's own email address, current date/time, allowed senders/recipients, send limits, and outgoing review rules. This clearly separates it from every sibling, which all deal with individual messages or events.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

It states exactly when to call it ('once at the start of a session'), the ordering constraint ('before list_messages or send_message'), and the reason ('so you only write to allowed recipients'). No other sibling is in scope here, so no alternative needs discussion.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint, idempotentHint and destructiveHint=false, so safety is covered. The description adds real behavioral context beyond them: sort order (newest first), that cancelled events are included and flagged by status, that attendee answers are present, and that it changes nothing. It stops short of describing result size or any limits, but for a zero-param read tool this is solid.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Three tight sentences: what it returns first, then how to use it and what sibling to prefer, then the exclusion and the return shape. Every clause carries information and nothing is repeated from the annotations or schema.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a zero-parameter list tool with no output schema, the description supplies the missing pieces an agent needs: scope, ordering, inclusion of cancelled items, and the return envelope { events }. Nothing required to call or interpret it is absent.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The tool takes no parameters, so there is nothing for the description to disambiguate and the baseline of 4 applies. The description correctly implies no filtering input is available, which is consistent with the empty schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource with scope: 'all meetings this mailbox created with create_event', plus ordering ('newest first') and content ('including cancelled ones (status) and attendee answers'). It also explicitly excludes invitations from others, so an agent can distinguish it from list_messages and get_event without opening a schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit when-to-use ('find an event id you lost'), names the sibling alternative for a different case ('use get_event for one event'), and states the exclusion plus where the excluded data lives ('Invitations from others are not listed here; they arrive as messages (see read_message)'). Routing is unambiguous.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare read-only, idempotent, non-destructive behavior, yet the description still adds material context: only mail from allowed senders is included, results are newest-first, the return shape is { messages, next_cursor }, and next_cursor round-trips as cursor. 'Changes nothing' reinforces the safety profile in plain language.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads purpose, then usage, then the alternative tool, then the return contract in a tight set of sentences. No redundant restatement of the name or schema.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Despite having no output schema, the description specifies the return shape and pagination. Combined with the sender-allowlist constraint and the harmless-read guarantee, an agent has everything needed to call and interpret this tool.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so parameters are documented structurally. The description still adds a layer of intent by grouping the filters into two use cases (unread check vs. older-mail search) and explaining the cursor round-trip, which is more than the schema alone conveys.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb ('list or search') and resource ('received mail in the inbox'), plus ordering and the returned fields. It also explicitly distinguishes itself from read_message, which returns the full text of one message.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives concrete when-to-use guidance: 'Use unread: true to check for new mail; text, from, subject, since and before to search older mail.' It names the alternative tool (read_message) and the condition that selects it.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint=false, idempotentHint=true, and destructiveHint=false, so the safe-mutation profile is covered structurally. The description adds genuinely new context: only the server-side read flag changes, nothing is sent, and the result is { ok: true } — useful for an agent reasoning about side effects.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences, zero filler, with the core action front-loaded and the scoping/return details trailing. Every clause earns its place.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a two-parameter mutation tool with full schema coverage and annotations covering the safety profile, the description supplies the remaining gaps: mutation scope, absence of outbound side effects, and the return shape (no output schema exists). Nothing needed to call it correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, and the schema already explains that 'id' comes from list_messages and that 'unread' means true=unread / false=read. The description adds no parameter syntax or format detail beyond that, so the baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('Mark a received message as read or unread') and immediately bounds the scope to the read flag only, which separates it from siblings like read_message, delete_message, and reply_message. An agent can identify the operation without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Provides a concrete use case ('put it back on your to-do pile after reading it with read_message'), which implicitly routes the agent: use read_message to read, this tool to flip the flag. There is no explicit when-not guidance or mention of alternatives like delete_message, so it stops short of a 5.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations declare readOnlyHint=false, and the description explains why: it marks the message as read. It also discloses the default-on side effect, the opt-out, the return format (JSON), and adds a prompt-injection caution. This goes well beyond the annotation surface.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

One tight paragraph that front-loads the purpose, then the payload, then routing, then the side effect and safety note. No sentence is filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

No output schema exists, yet the description enumerates the return contents so an agent knows what to expect. Combined with the side-effect disclosure and injection warning, nothing needed to call it correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3, but the description adds real meaning by tying mark_read's default to an observable side effect ('Marks the message as read unless mark_read is false') and clarifying the id source as coming from list_messages.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a precise verb+resource ('Read one received message') and enumerates exactly what is returned: headers, Markdown body, attachment list with index/name/type/size, and calendar invitation. This is clearly distinguishable from list_messages and the other siblings.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit sequencing ('Use it after list_messages'), names the follow-up tool for files ('fetch files with get_attachment'), and covers the conditional case via mark_read. Nothing about when to reach for this tool is left to inference.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only cover the generic safety profile (readOnly=false, destructive=false, openWorld=true, idempotent=false). The description adds real behavioral context beyond them: recipients and Re: subject are auto-filled, the original is not quoted, and the same allow-list/review/send-limit regime applies. It even discloses the return shape, which is non-trivial given no output schema.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the core action and dense with useful facts; nothing is padding. The single long first sentence with two nested parentheticals is slightly cramped to parse, which keeps it just short of ideal.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a mutation tool with no output schema, the description covers the pieces an agent actually needs: what gets auto-populated, what is not included, the send-limit/permission regime, and the return keys ({ message_id, warnings }). Attachment semantics are fully handled by the schema, so nothing material is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the baseline is 3. The description still adds meaning: it explains the reply_all side effect precisely ('also To and Cc, except you') and reinforces that body_markdown is not auto-quoted, both of which go slightly beyond the schema wording.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource (reply to a received message) and immediately scopes it (in the same thread). It names the sibling it supersedes ('Prefer this over send_message') and is clearly distinguishable from forward_message, so an agent can route without opening any schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives an explicit preference rule ('Prefer this over send_message for answers') plus the operating constraints that gate use: same allow list, review and send limit as send_message. Both when-to-use and the practical preconditions are stated rather than inferred.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only cover the safety profile (readOnlyHint=false, idempotentHint=false, destructiveHint=false, openWorldHint=true). The description goes well beyond them by disclosing the external side effect (a standard calendar reply is sent, the organizer's calendar updates), the authorization gate (organizer must be on allow_send_to), comment moderation, and the return shape { message_id, warnings }.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Three tight sentences with zero filler: capability first, routing/exclusion second, constraints and return shape last. Every clause carries information an agent needs.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a mutation tool with no output schema, the description covers the action, the side effect on the organizer, the authorization prerequisite, the alternative tool to avoid, and the returned fields. Nothing material is left for the agent to infer.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3. The description adds real meaning beyond the schema for `comment` (it is reviewed like any mail, i.e. subject to moderation) and clarifies where `id` comes from. It does not elaborate further on `response`, but the enum is self-explanatory.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb set (accept/decline/tentative) against a specific resource (calendar invitation you received), and explicitly distinguishes itself from send_message as the wrong path for invitations. An agent can identify the operation without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit when-to-use context ("read_message shows it under invitation") and an explicit when-not/alternative ("do not answer invitations with send_message"). The precondition that the organizer must be on allow_send_to is stated up front.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations declare readOnlyHint=false, openWorldHint=true, idempotentHint=false, and destructiveHint=false. The description adds valuable context beyond this: gateway review may reject with review_rejected and reasons, counts towards hourly send limit, saved to Sent, and returns { message_id, warnings }. It doesn't mention whether attachments are scanned or if review is asynchronous, but the disclosure is rich for a tool with annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Three sentences, front-loaded with purpose, then sibling routing, then behavioral constraints. Every sentence adds value; no redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

The description covers purpose, alternatives, allow_send_to precondition, review rejection, rate limit, storage, and return shape. With no output schema, it correctly includes the return object. However, it doesn't mention max attachment count (20) or body size limit (512000) – though those are in the schema. Minor gaps remain but overall complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the schema already documents all 7 parameters in detail. The description adds no parameter-specific syntax or format information beyond what the schema provides. Baseline 3 is appropriate when the schema does the heavy lifting.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific verb+resource ('Send a new email') and immediately specifies the format ('written in Markdown (sent as HTML plus plain text)'). It distinguishes itself from siblings 'reply_message' and 'forward_message' by name, so an agent can pick the right tool without opening schemas.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly states when to use alternatives: 'Use reply_message to answer a received message and forward_message to pass one on.' It also sets a precondition: 'Every recipient must be on allow_send_to.' This is clear when-to-use and when-not-to-use guidance.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.